After OpenAI's quasi-Riemann claim, human mathematicians prove 19-year-old conjecture without AI
新智元 · wechat · 2026-10-18
OpenAI claimed on Oct 6 that its AI proved the quasi-Riemann hypothesis (ruling out Landau–Siegel zeros, a 199-page proof), stunning number theorists. But humans struck back within days: Cureton & Thorner extended the AI-derived tools to finite-order Hecke L-functions, proving zero-free regions past 1-1/6000; a global open-source effort then pushed that boundary to 1/32, formalized in Lean by AxiomProver. Building on this, Frei, Somos and Widmer unconditionally proved the 2007 Ellenberg–Venkatesh ℓ-torsion bound for class groups of number fields — previously conditional on GRH and open for 19 years — ending their paper with the declaration: "No artificial intelligence tools were used in this work." Burgisser also proved Valiant's conjecture (VP≠VNP over ℂ) in the aftermath.
More from Models
- User Slams Claude's Slack Integration as Worst-in-class for Personal Accounts — akumar · 2026-10-19
- 4-year bitter-lesson evolution of MLLM vision: from LLaVA crops to native-resolution images — gabriberton · 2026-10-19
- DeepSeek V4.1 Flash matches Astra on 3D, but vision quality still lags rivals — bookwormengr · 2026-10-19
- Investor urges Google to distill Argon into a small model priced at 50% of Luna to top OpenRouter — bindureddy · 2026-10-19
- DeepSeek V4.1 Flash becomes first open model to crack top 10 on search leaderboard — gaganghotra_ · 2026-10-19
- KV Cache Compression Makes 1M Context Manageable at Under 100GB Per User — khademinori · 2026-10-19